In spite of numerous advances in medical research, cancer remains the second leading cause of death in the United States. Traditional modes of clinical care, such as surgical resection, radiotherapy and chemotherapy, have a significant failure rate, especially for solid tumors. Failure occurs either because the initial tumor is unresponsive, or because of recurrence due to regrowth at the original site or metastasis. Cancer remains a central focus for medical research and development.
Under the hypothesis that neoplastic cells are normally regulated by immune surveillance, an attractive approach is to re-focus the immune system in affected individuals back towards the tumor. Many types of cancers should be susceptible to the immune system, because they express unusual antigens that reflect the oncogenic transformation of the cell. Antibodies or T cells directed against a target antigen specifically expressed on tumor cells may be able to recruit immune effector functions, and thereby eliminate the tumor or mitigate the pathological consequences.
There are several potential pitfalls in this approach. The first is that the target tumor antigen may be shed from the cell, and thereby block the approach of tumor-specific immune components. The second is that the expression of the target antigen may be heterologous. In this case, a specific immune response would be ineffective in a substantial subclass of affected individuals, or the risk of escape variants would be high. The third difficulty is that tumor-specific antigens are generally antigenically related to autoantigens, or comprise autoantigens in an atypical mode of expression. This means that tumor-associated antigens are often poorly immunogenic; perhaps due to an active and ongoing immunosuppression against them. Furthermore, cancer patients tend to be immunosuppressed, and only respond to certain T-dependent antigens.
A number of features suggest that gangliosides may be preferable to other types of target antigens for antibody-mediated killing of certain tumor types. Gangliosides like GD2 have simple, well-defined structures, and the level of expression is not affected by antibody binding. In vitro studies have shown that monoclonal antibodies against gangliosides like GD2 and GD3 potentiate lymphocyte response which could potentially be directed towards tumor cells. In addition, certain gangliosides have been implicated in the adhesion of tumor cells to solid substrates, and antibodies inhibit this attachment. An immune response against them, even if not successful in eliminating the tumor cells, could reduce the extent of pathology.
In particular, glycosphingolipid GD2 is expressed at high density by tumors of human neuroectodermal origin; including malignant melanoma, neuroblastoma, glioma, soft tissue sarcoma and small cell carcinoma of the lung. The GD2 antigen is absent in most normal tissues, except for low levels in brain and peripheral nerve.
Melanoma is one of the human diseases for which there is an acute need of new therapeutic modalities. It is a particularly aggressive form of skin cancer, and occurs in increased frequency in individuals with regular unguarded sun exposure. In the early phases, melanoma is characterized by proliferation at the dermal-epidermal junction, which soon invades adjacent tissue and metastasizes widely. Worldwide, 70,000 patients are diagnosed and 25,000 deaths are reported each year. The American Cancer Society projects that by the year 2000, 1 out of every 75 Americans will be diagnosed with melanoma in their lifetime.
Fortunately, melanoma is one of the cancers for which gangliosides hold significant promise as a target antigen (Livingston (1995) Immunol. Rev. 145:147-166). Increased expression of GD2 has been observed in a majority of malignant melanoma cells. Several murine monoclonal anti-GD2 antibodies were reported to suppress the growth of tumors of neuroectodermal origin in athymic (nu/nu) mice or cause remission in patients with metastatic melanoma. A human-mouse chimeric anti-GD2 antibody remissions in patients with metastatic neuroblastoma. The mechanism is thought to involve antibody dependent cellular cytotoxicity (ADCC) or complement-mediated cytotoxicity (CMC). Clinical responses have been obtained by treating with monoclonal antibodies against GM2, GD2 and GD3. Active immunization with a ganglioside vaccine comprising GM2 produced anti-GM2 antibodies in 50/58 patients, who survived longer on average than antibody negative patients.
Neuroblastoma is a highly malignant tumor occurring during infancy and early childhood. Except for Wilm's tumor, it is the most common retroperitoneal tumor in children. This tumor metastasizes early, with widespread involvement of lymph nodes, liver, bone, lung, and marrow. While the primary tumor is resolvable by resection, the recurrence rate is high.
Small cell lung cancer is the most malignant and fastest growing form of lung cancer. It accounts for 20-25% of new cases of lung cancer, and 60,000 cases will be diagnosed in the U.S. in 1996. The primary tumor is generally responsive to chemotherapy, but is shortly followed by wide-spread metastasis. The median survival time at diagnosis is .about.1 year, with a 5 year survival rate of only 5-10%.
If there was a simple and reliable therapeutic strategy for providing immune reactivity against GD2, then the clinical prospects for these types of cancers might improve.
Unfortunately, there are several reasons why GD2 is less than ideal as a component of an active vaccine. For one thing, GD2 is of limited supply, and is difficult to purify. Of course, because GD2 is a ganglioside, it cannot be generated by simple recombinant techniques. Secondly, gangliosides in general, and GD2 in particular, are poorly immunogenic. In order to render them more immunogenic in humans, it has been necessary to conjugate them to protein carriers like KLH (U.S. Pat. No. 5,308,614; WO 94/16731).
Similarly, the passive administration of anti-GD2 antibodies is less than ideal as an approach to long-term care. The amount of antibody that must be provided passively is substantial. It may lead to the formation of anti-immunoglobulin or anti-idiotype, which in turn may lead to diminished responsiveness with each succeeding dose. And perhaps most importantly, it fails to provide the host with components of the immune response which may be critical in tumor eradication; particularly tumor-directed cytotoxic T cells.
How else, then, could an active immune response against GD2 be obtained? The network hypothesis of Lindemann and Jerne suggests a way of overcoming both the natural immune tolerance against GD2, and the shortage of supply of GD2. It relies on the fact that antibodies comprise variable region epitopes that themselves may be immunogenic, leading to the generation of second-level antibodies called anti-idiotypes. According to the hypothesis, immunization with a given tumor-associated antigen will generate production of antibodies against this tumor-associated antigen (Ab1), which is purified and used in a second round of immunization to generate the anti-idiotype (Ab2). Some of these Ab2 molecules (called Ab2.beta.) fit into the paratopes of Ab1, and can be used as surrogate tumor-associated antigens. Thus, immunization with Ab2.beta. can lead to the generation of anti-anti-idiotype antibodies (Ab3) that recognize the corresponding original tumor-associated antigen identified by Ab1. The idea is that the Ab2 presents a feature that is structurally related to the tumor antigen in a different context that makes it more immunogenic.
Accordingly, efforts have been made elsewhere using anti-idiotypes to elicit a response against various tumor-associated antigens. One group of investigators raised an anti-idiotype related to the melanoma associated ganglioside GM.sub.3 (Kanda et al., Yamamoto et al., Hastings et al.). Saleh et al. and Cheung et al. have raised anti-idiotypes against GD2. Other anti-idiotypes have entered early clinical trials: for example, Mittelman et al. are using an anti-idiotype related to a high molecular weight melanoma associated antigen (MAA). Chattopadhyay et al. have developed anti-idiotypes against a melanoma-associated proteoglycan. Chapman et al are treating melanoma patients with an anti-idiotype related to ganglioside GD3. Most of these trials are still at about the phase I level, with results still pending.
It is important to emphasize that every potential anti-idiotype for use in tumor therapy must be evaluated on a case-by-case basis. First, only a fraction of antibodies raised against an Ab1 are limited in their reactivity against the paratope of the Ab1 (i.e., non-reactive against features shared with other potential antibodies in the host individual). Second, anti-idiotypes are not necessarily immunogenic. Third, only a fraction of immunogenic anti-idiotypes elicit a response against the original tumor antigen and not against other antigens with less tissue specificity. These properties are related to the structure of the anti-idiotype, including the amino acid sequence of its variable regions. Different anti-idiotypes raised against the same Ab1 will be different in this respect, and must be evaluated separately. Furthermore, the anti-tumor response elicited will depend on the immunological status of the individual being immunized.